Reducing 4G/LTE testing time with trial acceleration

The GSA (Global mobile Suppliers Association) recently published an update to its Evolution to LTE report. The findings: more than 200 operators are investing in LTE, with 154 LTE network deployments in progress or planned in 60 countries, including 20 networks which have commercially launched. A further 54 operators in 20 additional countries, according to the GSA, are engaged in LTE technology pilot trials or tests ahead of formal commitments to deploy commercial networks.

The above numbers don’t include the WiMAX, WiMAN and/or WiBro tests underway by Intel, KDDI, KT, LG, Intel, Sprint, Samsung and others. Complicating this further is the adoption of Voice over LTE (VoLTE), for which standards are still developing, such as the GSMA’s IR.92/OneVoice recommendation. With an estimated more than 140 4G trials of all kinds underway, it’s safe to say there’s a whole lot of testing going on.

LTE trials usually involve a lot of repetitive cabling/uncabling and configuration of both the test network and the device under test itself. Very often there is more than one device in the test, each device needing to be swapped in and out multiple times. Many test protocols call for serialised testing, whereby dozens or hundreds of steps must be followed in order, the better to simulate real-world interoperability conditions. And interoperability can be complex given the need to work well with 2G/3G, whether GSM or CDMA, and with analog systems.

In talking with those implementing tests at large carriers and listening to others at industry meetings, I’ve heard that more than 50% of testing time is consumed by these standard operations as opposed to the actual testing itself. This has been the nature of the beast for quite some time for trials and evaluations.

Trial acceleration

More recently there has been a significant advance in 4G telecoms testing for how trials and evaluations. The new method is called ‘trial acceleration’. It helps service providers to get through the evaluation phase and to market faster, and it allows much more control when working with existing products and new software loads.

The main breakthrough comes in minimising wiring/rewiring over and over when configuring equipment and the test network. Wiring needs to be done just once per physical test scenario. Timeline gains of up to 60% are not uncommon - according to one leading worldwide mobile carrier, mainly due to the heavily repetitive nature of 4G trial/evaluation testing.

Another problem in trialling is that of trace capture and analysis. This brings into consideration the issues with switch SPAN ports, switches and hubs as a means of accessing the network for packet capture.

Here is a test case for LTE with repeated swapping-in or swapping-out of multiple same-function network elements, such as the P or S Gateways, the Policy and Charging Rules Function or the Mobility Management Entity function as well as logical elements within the IP Multimedia Subsystem core.

The Trial Acceleration System is, in many ways, similar to a switch or router, but with several additional capabilities. Unlike a switch or router, the system acts as a network of intelligent TAP points connected as a system, sitting inline between the mobile core and the system under test. Therefore, the monitoring system has access to all levels of communication - from OSI Layer 2 upwards - overcoming many of the shortcomings of switches, their lack of visibility in to a test scenario and the way that SPAN or mirror ports function in masking Layer 2 information such as jitter.

Initial connection to the network

The trial acceleration system takes the place of the device or devices under test. All wiring is carried out as usual, but terminates in the trial acceleration system instead of multiple wirings to several similar or different network equipment. Bidirectional traffic flows though the trial acceleration system between the core and the specific device under test at that time for that scenario. Results are logged and traces can be collected. All traffic is made available to a special monitoring port whereby analysers or monitoring tools can be connected.

In practice, the trial acceleration system is reconfigured through a management GUI for each test run - with each network element that is not currently involved with the testing being isolated and unable to see any third-party traffic/packets. With several dedicated monitoring ports built in, trials can now be carried out more comprehensively and with greater depth of monitoring.

Benefits include:

• More in-depth reporting with greater monitored network and interface coverage.
• Trace the path of a packet through the network end-to-end.
• Minimise manual equipment configuration.
• Reduce switch SPAN/mirror port contention.
• Effectively use Layer 2 information such as jitter and latency in diagnosing potential QoS/QoE issues.
• Speed up fault-finding when systems are not working as expected.
• Monitor all network nodes/interfaces at the same time.
• Attach multiple analyser devices at the same time.

Given how much of testing time is devoted to physical connectivity tasks, techniques like trial acceleration can help complete carriers complete tests faster.

By Andy Huckridge, Director of Marketing, VSS Monitoring